Multiple Object Motion Detection for Robust Image Stabilization Using Block-Based Hough Transform

Abstract

In this paper, a new digital image stabilization involving multiple objects motion detection based on block-based Hough transform (HT) is presented. The line edges in a pair of edge blocks, where one is from current frame and the other is from previous frame are aligned to generate potential motion vectors to have votes on parameter voting matrix H using modified HT. The parameter voting matrix H is used for identifying feasible motion vectors, which generate the representative motion vector for compensation. The unexpected camera vibration is removed regardless of object's movement using the proposed image stabilization algorithm. The experimental results show that the proposed method outperforms traditional block-matching based digital image stabilization methods.